release: v0.26.0 timeline and caption workflow upgrades

[renezander030]

Summary

• Add exact frame-grid linting and boundary-derived repairs.
• Improve Chinese/Japanese script alignment and add a minimum-match write gate.
• Select explicit audio streams before caption transcription.
• Add boundary-safe ripple deletion and partial timeline shifts.
• Move large FFmpeg filter graphs into temporary script files.
• Flatten nested OpenTimelineIO timelines while preserving offsets and reporting unsupported data.
• Add CRF and target-bitrate proxy render controls.
• Add progressive word-reveal captions.
• Add atomic preset-based restyling for whole text tracks.

This prepares the coordinated capcut-cli 0.26.0 and Python capcut 0.1.1 release, including regenerated command references and English/Chinese release notes.

Verification

• npm test — 1,000 passed
• npm run lint
• python3 -m unittest discover -s python/tests -v — 11 passed
• npm pack --dry-run --json
• Python wheel build for capcut==0.1.1
• Generated command-reference JSON parses successfully

Publishing gate

No package has been published. After review approval, release targets are npm (capcut-cli@0.26.0) and PyPI (capcut==0.1.1). MCP Registry publication does not apply because this repository does not ship a standalone MCP server; its experimental Wasm component is hosted through Wassette.